Policy : Run to Stop When Problems Arise

Commitment to Fast, Safe Incident Response
We believe that when something breaks, our priority is not to press on—it’s to pause, respond, and recover with discipline. The sooner we stop unsafe work, the sooner we can restore stability and learn. Reacting quickly and calmly to incidents protects users, systems, and teams alike.

What This Means
We empower every engineer to halt progress when safety is at risk. Our response to incidents prioritises system improvement, team support, and organisational learning over blame or speed. Pausing is a strength, not a failure.

Our commitment to running to stop is built on:

• Empowered Halts and Rollbacks – Any engineer can trigger a rollback or pause a release if they believe an issue threatens safety or performance. Clear escalation paths exist, and team support is immediate.
• Actionable, Calibrated Alerting – Alerts are designed for clarity and urgency, not noise. We actively reduce alert fatigue by ensuring all alerts are actionable and meaningful.
• Supported On-Call Practices – Engineers on-call have access to documented runbooks, auto-remediation tools, and responsive team support. They are never left to firefight alone.
• Incident Follow-Up Leads to Change – Incident handling results in meaningful platform or service improvement—not just postmortem paperwork. RCA is a starting point, not the end.
• Blameless, Timely Reviews – Post-incident reviews are conducted promptly, include all relevant functions, and focus on learning, not blame. We ask what went wrong in the system, not who made the mistake.

Why This Matters
Rushing through problems increases risk and erodes trust. Slow or disempowered responses to incidents cause harm to users and burnout for teams. Running to stop enables safe systems, healthier engineers, and a culture that treats learning as a first-class activity.

Our Expectation
All teams must operate with the mindset that stopping is responsible, not disruptive. Incident protocols, runbooks, and alert hygiene must be maintained. Leaders must back the decision to pause work and invest in prevention, not just recovery.

To support this policy, teams will be guided by standards for incident response, alert design, on-call health, and continuous improvement from post-incident reviews. By running to stop when problems arise, we protect what matters most—and build a culture of safety over speed.